[PATCH] Fix PR target/6540

Thu May 2 13:05:00 GMT 2002

   From: Richard Henderson <rth@redhat.com>
   Date: Thu, 2 May 2002 13:00:35 -0700

   On Thu, May 02, 2002 at 02:38:34PM -0400, Jakub Jelinek wrote:
   > But if e.g. Solaris (and hopefully FreeBSD) defines __LONG_DOUBLE_128__
   > unconditionally, why shouldn't it use float-i128.h then?

   I agree.

You really think the maintenance argument holds no merit?

If so, I'll withdraw my patch and we can go with something
more like Jakub's scheme to fix this bug.

But I really think there is value in centralizing all of this
stuff.